Electrical signals recorded from the maternal abdomen can in principle be used to monitor the electrical activity of the fetal heart from early in pregnancy through delivery. In practice the fetal EKG signal is often so small, relative to the interferring electrical noise, that it is very difficult to detect. We propose to develop an adaptive signal processing system to improve the signal to noise ratio in both direct and indirect fetal electrocardiography. The system would eliminate the maternal EKG signal from the fetal record, eliminate 60 Hz interference, and permit coherent spatial summation of multiple abdominal leads for further improvement of the signal to noise ratio. The fetal EKG signal would be preserved in beat-to-beat detail. Our work will proceed in three stages: 1) we will develop a set of optimized adaptive algorithms for fetal EKG processing; 2) we will develop the hardware and software required to realize these algorithms; 3) we will clinially evaluate the effectiveness of the signal processing system.